Regardless of who has the superior algorithm, heavy lifting tasks can be performed much more efficiently and with much higher precision on chips and FPGAs. This is a well known fact....

[Jud]

Same thing here: I end up using a RamDisk creator for Audirvana+ and launch it from there.

 

 

- I use a script that runs on bootup to (1) load A+ to a RAMdisk; (2) load music files to a folder in another RAMdisk; I have A+ set up with a smart playlist that will include any files in this folder. So A+ in a RAMdisk is ready to play music from another RAMdisk when my computer boots. This could also be used for HQPlayer. AppleScript, no external software necessary. Be happy to post the script if you would like.

 

- As of right now, apps won't run from a RAMdisk in the latest El Capitan beta, absent some magic I haven't come across yet. (Haven't had time to read up and see whether it's a known issue and what solution might exist.) So if you want to run your player apps from RAMdisk, stay with Yosemite for the time being.

[YashN]

But I recommend using iFi iUSB Power, it seems to improve (measured) performance.

 

Interesting.

[Blizzard]

Blizzard, I think you have a misunderstanding of what we're trying to tell you. It isn't that the NAA doesn't isolate the server from the DAC. That is its prime function, it does exactly that.

 

What we're saying is that in addition to that isolation, there are other phenomena that need to be taken into account. Other types of noise using alternate routes.

 

 

 

What precisely is your understanding of 'better algo' here?

 

Less complexity can be helpful if it allows a smaller noise profile, and smaller EMC issues.

 

But if you have what looks like a smaller, less complex implementation but the EMC/Power consumption/Power pollutin issues are bigger than a client-server with HQ Player + NAA does, your smaller implementation may well be less good regarding SQ.

 

 

 

No, some people are missing this, but what's most will still miss is it doesn't stop there at all. Implementation with EMC concerns and System integration are fundamental.

 

Have you tried at least HQ Player in direct mode from computer to DAC via USB (assuming you have USB async)?

 

Yes I have tried it. Like I've said many times already, I'm not going to be able to beat the SD card with any device that interfaces with the DAC via USB. If Miska offline converts some tracks for me to load on the SD card, it will be the ultimate test of his modulators/filters. I want to demo his technology to the Resonessence guys. Even if HQplayer sounds better than lets say Jriver on computer, the SD card still has an unfair advantage.

[Jud]

How do you know until different noise profiles in various implementations are actually put to the final test of SQ?

 

He has coffee with smart people.

[Blizzard]

You could try with Audiogate v2 or with Yuri's software.

 

It will be hard to test Miska's modulators/filters using someone else's software.

[YashN]

Sounds like spooky action at a distance to me!

 

Hehe, it does, doesn't it?

 

But it all has normal, rational explanations if you dig deep enough.

[YashN]

Sure, I would love to see DACs on the market that don't do any DSP, only D/A conversion very well.

 

Lampizator chipless DSD DACs?

[Blizzard]

How do you know until different noise profiles in various implementations are actually put to the final test of SQ?

 

How can putting more electronics into the chain possibly reduce noise/jitter. That's like saying if you are using super good speaker cable, instead of using 10 feet use 100 because it's really good so more must be better.

[YashN]

I personally don't like sound of Sabre's modulators...

 

I used to read a lot about the ESS Sabre chips and DACs. In the end, getting Burr-Brown ones helped me forget them

 

Burr-Brown sounds sweet!

[Blizzard]

Interesting.

 

This is because the Amanero is powered by the USB cable. Without the Ifi, it uses the dirty power from the PC.

[Blizzard]

I used to read a lot about the ESS Sabre chips and DACs. In the end, getting Burr-Brown ones helped me forget them

 

Burr-Brown sounds sweet!

 

It's all about implementation. There's much more to a DAC than the chip.

[Paul R]

Why do you assume this is the case?

 

Uh, because not only is the NAA logically isolated from the server, but it is physically isolated from the server as well.

 

 

 

Nope, EMC and noise matters everywhere in the system, because there are alternate routes (e.g. the AC question where the server could be polluting the mains and then from there this affects your DAC and amp, the amp could also affect the DAC, etc...).

 

And then there's the question of ground loops... and more.

 

Not so! A server can be spraying EMI all around it and throughout the internal case, but it will not affect the NAA at the other end of a network, isolated with intervening walls, switches, cables, and so forth.

 

The only things that can affect the sound coming out of the NAA are local physical parameters. Such as *local* power, and other such factors in the *local* environment.

 

The server and the NAA, or actually, any server and any remote low power low noise player, are quite distinct entities, and that is on purpose.

 

-Paul

[Paul R]

How can putting more electronics into the chain possibly reduce noise/jitter. That's like saying if you are using super good speaker cable, instead of using 10 feet use 100 because it's really good so more must be better.

 

Well duh- what do you think a reclocker does? ASRC? Alex and John's REGEN? All of which serve to reduce jitter going into the DAC.

 

Jitter, unless it is so sever as to cause data errors, really is a non issue until it hits the DAC's conversion section.

 

In your particular case, jitter, or lack thereof, from an SD card transmission can possibly affect the sound. But many find that just using an ethernet connected disk does the same thing, sometimes even better. No electrical connection there at all...

[YashN]

Typically an OS memory maps the program file which is loaded into RAM via pagefaults -- when a memory loc is accessed that is not yet in memory, that page is loaded. I can see the possibility at the beginning of running a program, that there may be a difference e.g. if pagefault to SSD generates some SSD signal whereas to RAMdisk not. But I don't understand how this could be the case once the program has been running for some seconds.

 

I've wondered if entirely eliminating *any* disk would be even better. This can be accomplished using iSCSI boot. But still I'm stumped as to why this could be the case? Any kernel hackers with ideas?

 

Any resources in contention are bound to significantly increase the noise profiles. There are a few threads where people mention booting from SD Cards, and then completely disconnecting or disabling HDDs/SSDs.

 

More than that I'm also stumped as to how this effect could pass through an optical cable and 2 FMCs which should reclock the signal. Stumped.

 

Ha, but who said it passes here?

[Blizzard]

Well duh- what do you think a reclocker does? ASRC? Alex and John's REGEN? All of which serve to reduce jitter going into the DAC.

 

Jitter, unless it is so sever as to cause data errors, really is a non issue until it hits the DAC's conversion section.

 

In your particular case, jitter, or lack thereof, from an SD card transmission can possibly affect the sound. But many find that just using an ethernet connected disk does the same thing, sometimes even better. No electrical connection there at all...

 

If computer/renderer combos are completely transparent, you should be able to daisy chain 20 of them before the DAC and get the same results.

[YashN]

Checkout Mutec for some pro level reclockers. My setup used to sound different depending on whatever, but now it sounds the same regardless of whatever.

 

Interesting feedback, ilok, after the one on Balanced Power. I am curious which Mutec device you used and where, and generally, your whole system implementation if you can tell me more.

[YashN]

This leads me to believe that the stability of the USB clock will indeed have a audible SQ difference

 

There is a lot of complex analog circuitry which is reliably giving stable digital results in numeric value domain, but the consistency of the timing of delivery seems to be a factor. Perhaps this inconsistent timing is causing additional electrical noise generation further along the chain or their could be other effects. As lmitche ​said to me the other day "Empiricism Rules"

 

On another note, Win 10 and Win 8 both sound significantly better when the HQ Player execution priority is changed from normal to realtime, and this is still the case when using NAA. Further evidence for the importance of timing.

 

Not just timing, the very operation of any digital sub-system has an effect, including hardware and software (the memory used, the algs, the buffer architecture used, etc...)

[YashN]

No need to be stumped. It just is so - always has been. And not even the wildest means of isolation helps.

Just utilize the "knowledge" ...

 

I am running the audio PC without any device in there, since 5 weeks or so (Windows 10). Not even an SD card. And yes, it again matters.

CPU usage is virtually zero (32/705600 output).

 

Long time no hear, Peter, what's up? Had a great holiday?

 

Anything new of note in XXHighEnd + Phasure NOS1a land?

[YashN]

I use a script that runs on bootup to (1) load A+ to a RAMdisk; (2) load music files to a folder in another RAMdisk; I have A+ set up with a smart playlist that will include any files in this folder. So A+ in a RAMdisk is ready to play music from another RAMdisk when my computer boots. This could also be used for HQPlayer. AppleScript, no external software necessary. Be happy to post the script if you would like.

 

That's a much better way to do it than what I'm currently doing. Never played with AppleScript, but I really should, or else 'll just ping you, Jud, thanks

 

I did also try using another RAMdisk for the files, but since I usually like to listen to DSD128 and to whole albums at one go, 4Gbs of DRAM turns up to be very small, so currently, out of convenience, I still let A+ read from the external HDD which is less than ideal.

 

As of right now, apps won't run from a RAMdisk in the latest El Capitan beta, absent some magic I haven't come across yet. (Haven't had time to read up and see whether it's a known issue and what solution might exist.) So if you want to run your player apps from RAMdisk, stay with Yosemite for the time being.

 

I'm actually writing from the Mavericks iMac, since my Yosemite MBP has gone AOL, internal HDD won't boot, probably that recurrent over-heating issue.

 

I wasn't impressed with the Yosemite beta on the MBP, it made everything super slow as well as showed a lot of Wi-Fi network issues, things that were reported by a lot of people including me as making OS X impossible to use normally, but which weren't even fixed for several patch releases. Very disappointing.

 

So, for now, we really have no desire to try El Capitan especially in beta.

[YashN]

He has coffee with smart people.

 

Sure helps a lot of conjecturing.

[YashN]

It will be hard to test Miska's modulators/filters using someone else's software.

 

At least it would allow you to test what an offline converted file in your pristine SD Card system does without the main server computer in the picture as you seem to abhor, doesn't it?

[YashN]

How can putting more electronics into the chain possibly reduce noise/jitter. That's like saying if you are using super good speaker cable, instead of using 10 feet use 100 because it's really good so more must be better.

 

No, it's all about system implementation.

 

You could have a tiny device in electronic complexity terms and it could still sound like crap compared to a bigger one because of various things.

 

Generally, I do favour lower complexity circuits and appropriate software as I want a lower noise profile. Hence the search for a real audiophile motherboard and a proper, simple NAA, because having a computer in the loop for me provides more flexibility for what I want to build in the future and also the networking/streaming flexibility.

[YashN]

This is because the Amanero is powered by the USB cable. Without the Ifi, it uses the dirty power from the PC.

 

I already know.

[Blizzard]

At least it would allow you to test what an offline converted file in your pristine SD Card system does without the main server computer in the picture as you seem to abhor, doesn't it?

 

Yes that's the point. Then I'll have a true reference point. Next step will be to try to match the performance using HQ player and the same modulator/filter combo.

[YashN]

Uh, because not only is the NAA logically isolated from the server, but it is physically isolated from the server as well.

 

And you think that's all that matters?

 

Not so! A server can be spraying EMI all around it and throughout the internal case, but it will not affect the NAA at the other end of a network, isolated with intervening walls, switches, cables, and so forth.

 

Well, you could start by not replacing a crappy shortcut by another crappy shortcut, in particular, stop using 'spraying' as if the only thing that happening is like how an aerosol works. Secondly, where's the DAC in your paragraph, and optional pre-amp and amp?

 

The only things that can affect the sound coming out of the NAA are local physical parameters. Such as *local* power, and other such factors in the *local* environment.

 

The server and the NAA, or actually, any server and any remote low power low noise player, are quite distinct entities, and that is on purpose.

 

Think more about power and ground loops and you'll have a bigger understanding.